Inkjet printers are now commonplace in homes and offices. For example, inkjet photographic printers, which print color images generated on digital cameras, are, to an increasing extent, replacing traditional development of photographic negatives. With the increasing use of inkjet printers, the demands of such printers in terms of print quality and speed, continue to increase.
All commercially available inkjet printers use a scanning printhead, which traverses across a stationary print medium. After each sweep of the printhead, the print medium incrementally advances ready for the next line(s) of printing. Such printers are inherently slow and are becoming unable to meet the needs of current demands of inkjet printers.
The present Applicant has previously described many different types of pagewidth printheads, which are fabricated using MEMS technology. In pagewidth printing, the print medium is continuously fed past a stationary printhead, thereby allowing high-speed printing at, for example, one page per 1-2 seconds. Moreover, MEMS fabrication of the printhead allows a much higher nozzle density than traditional scanning printheads, and print resolutions of 1600 dpi are possible.
Some of the Applicant's MEMS pagewidth printheads are described in the patents and patent applications listed in the cross-references section above, the contents of which are herein incorporated by reference.
To a large extent, pagewidth printing has been made possible by reducing the total energy required to fire each ink droplet and/or efficiently removing heat from the printhead via ejected ink. In these ways, self-cooling of the printhead can be achieved, which enables a pagewidth printhead having a high nozzle density to operate without overheating.
However, whilst a total amount of energy to print, say, a full-color photographic page will be approximately constant for any given pagewidth printhead, the power requirement of the printhead may, of course, vary. An average power requirement for printing a page is determined by the total energy required and the total time taken to print the page, assuming an equal distribution of printing over the time period. In addition, the power requirement of the printhead during printing of the page may fluctuate. Due to a particular configuration of the printhead or printer controller, some lines of print may consume more power than other lines of print. Hence, a peak power requirement for each line of printing may be different.
In a typical pagewidth printhead, nozzles ejecting the same color of ink are arranged longitudinally in color channels along the length of the printhead. Each color channel may comprise one or more rows of nozzles, all ejecting the same colored ink. In a simple example, there may be one cyan row of nozzles, one magenta row of nozzles and one yellow row of nozzles. Usually, each row of nozzles will be fired sequentially during printing e.g. cyan then magenta then yellow.
Furthermore, a typical pagewidth printhead may be comprised of a plurality of printhead modules, which abut each other and cooperate to form a printhead extending across a width of the page to be printed. Each printhead module is typically a printhead integrated circuit comprising nozzles and drive circuitry for firing the nozzles. The rows of nozzles extend over the plurality of printhead modules, with each printhead module including a respective segment of each nozzle row.
In previous patent applications, listed below, we described various types of printheads, printer controllers and methods of printing. The contents of these patent applications are herein incorporated by reference:
10/854,52110/854,52210/854,48810/854,48710/854,50310/854,50410/854,50910/854,51010/854,49610/854,49710/854,49510/854,49810/854,51110/854,51210/854,52510/854,52610/854,51610/854,50810/854,50710/854,51510/854,50610/854,50510/854,49310/854,49410/854,48910/854,49010/854,49210/854,49110/854,52810/854,52310/854,52710/854,52410/854,52010/854,51410/854,51910/854,51310/854,49910/854,50110/854,50010/854,50210/854,51810/854,51710/934,62811/212,823
In our previous patent applications U.S. Ser. No. 10/854,498, filed May 27, 2004, U.S. Ser. No. 10/854,516, filed May 27, 2004 and U.S. Ser. No. 10/854,508, filed May 27, 2004, we described a method of printing a line of dots where not all nozzles in one row or one segment are fired simultaneously. Rather, the nozzles are fired sequentially in firing groups in order to minimize the peak power requirement during printing of one line. As a consequence, each line of printing is typically not a perfectly straight line (unless the physical arrangements of the nozzles directly compensates for the firing order in which case it can be a straight line), although this imperfection is undetectable to the human eye. Each segment on a printhead module may comprise, for example, 10 firing groups of nozzles, in order to minimize, as far as possible within the print speed requirements, the peak power requirement for firing that segment of the nozzle row.
In our previous patent applications U.S. Ser. No. 10/854,512, filed May 27, 2004 and U.S. Ser. No. 10/854,491, filed May 27, 2004, we described a means for joining abutting printhead modules such that the effective distance between adjacent nozzles (‘nozzle pitch’) in the row remains constant. At one end of each printhead module, there is a displaced nozzle row portion, which is not aligned with its corresponding nozzle row. The firing of these displaced nozzles is timed so that they effectively print onto the same line as the row to which they correspond. As such, all references to “rows”, “rows of nozzles” or “nozzle rows” herein include nozzle rows comprising one or more displaced row portions, as described in U.S. Ser. No. 10/854,512, filed May 27, 2004 and U.S. Ser. No. 10/854,491, filed May 27, 2004.
In our previous patent applications U.S. Ser. No. 10/854,507, filed May 27, 2004 and U.S. Ser. No. 10/854,523, filed May 27, 2004, we described a means by which the visual effect of defective nozzles is reduced. The printhead described comprises one or more ‘redundant’ color channels, so that for a first row of nozzles ejecting a given color, there is a corresponding second (‘redundant’) row of nozzles from a different color channel which eject the same color. As described in U.S. Ser. No. 10/854,507, filed May 27, 2004 and U.S. Ser. No. 10/854,523, filed May 27, 2004, one line may be printed by the first nozzle row and the next line is printed by the second nozzle row so that the first and second nozzle rows print alternate lines on the page. Thus, if there are unknown defective nozzles in a given row, the visual effect on the page is halved, because only every other line is printed using that row of nozzles.
Alternatively, if there are known dead nozzles in a given row, the corresponding row of nozzles may be used to print dots in those positions where there is a known dead nozzle. In other words, only a small number of nozzles in the ‘redundant’ row may be used to print.
As already mentioned, the redundancy scheme described in U.S. Ser. No. 10/854,507, filed May 27, 2004 and U.S. Ser. No. 10/854,523, filed May 27, 2004 has the advantage of reducing the visual impact of dead nozzles, either known or unknown. Moreover, careful choice of redundant colors may be used to further reduce the visual impact of dead nozzles. For example, since yellow makes the lowest contribution (11%) to luminance, the human eye is least sensitive to missing yellow dots and, therefore, yellow would be a poor choice for a redundant color. On the other hand, black, makes a much higher contribution to luminance and would be a good choice for a redundant color.
However, while the redundancy scheme described in U.S. Ser. No. 10/854,507, filed May 27, 2004 and U.S. Ser. No. 10/854,523, filed May 27, 2004 can compensate for dead nozzles and reduce (e.g. halve) the number of dots fired by some nozzles, it places increased demands on the power supply which is used to power the printhead. The reason is because in the time it takes for the print medium to advance by one line (one ‘line-time’), each nozzle row must be allotted a portion of the line-time in which to fire, in order to achieve dot-on-dot printing and provide the desired image. Each nozzle row is allotted a portion of the line-time, since not all nozzle rows can fire simultaneously. (If all nozzle rows were to fire simultaneously, there would be an unacceptable current overload of the printhead).
In a simple CMY pagewidth printhead, having three rows of nozzles and no redundant color channels, each nozzle row must fire in one-third of the line-time. If the average power requirement of the printhead is x, then the peak power requirement over the duration of the line-time is as shown in Table 1:
TABLE 1ColorPeak PowerLine-timeChannelRequirement0Cx0.33Mx0.67Yx0 (new line)Cx. . . etc.
In this simple CMY printhead with no redundant nozzles, power is distributed evenly over the duration of the line-time so that the peak power requirement is constant and equal to the average power requirement of the printhead. From the standpoint of the power supply, this situation is optimal, but, on the other hand, there is no means for minimizing the visual effects of dead nozzles.
In a CMY printhead having redundant cyan and magenta color channels (i.e. C1, C2, M1, M2 and Y color channels) and a pair of nozzle rows in each color channel (for even and odd dots), each nozzle row is allotted one-tenth of the line-time, since there are now ten nozzle rows. Now if the average power requirement of the printhead is x, with the redundancy scheme and firing sequence described in U.S. Ser. No. 10/854,507, filed May 27, 2004 and U.S. Ser. No. 10/854,523, filed May 27, 2004, the peak power requirement over the duration of two line-times is as shown in Table 2:
TABLE 2ColorPeak PowerLine-timeChannelRequirement0C11.67x(even)0.1C20(even)0.2M11.67x(even)0.3M20(even)0.4Y (even)1.67x0.5C1 (odd)1.67x0.6C2 (odd)00.7M11.67x(odd)0.8M20(odd)0.9Y (odd)1.67x0 (new line)C10(even)0.1C21.67x(even)0.2M10(even)0.3M21.67x(even)0.4Y (even)1.67x0.5C1 (odd)00.6C2 (odd)1.67x0.7M10(odd)0.8M21.67x(odd)0.9Y (odd)1.67x0 (new line)C11.67x . . . etc(even)
It is evident from the above table that the peak power requirement of the printhead fluctuates severely between 1.67x and 0 within the period of a line-time, even though the average power consumed over the whole line-time is still x. In practical terms, it is difficult to manufacture a power supply which is able to deliver severely fluctuating amounts of power within each line-time. Hence, the redundancy described in U.S. Ser. No. 10/854,507, filed May 27, 2004 and U.S. Ser. No. 10/854,523, filed May 27, 2004 is difficult to implement in practice, even though it offers considerable advantages in terms of reducing the visual effects of known dead nozzles.
Of course, a printhead could be configured not to fire redundant color channels in a given line-time, resulting in an average of x peak power for each nozzle row. Such a configuration is effectively the same as that described in Table 1. While this configuration would address peak power and misdirectionality issues, it would not address the problem of known dead nozzles, since only one of each redundant color channel would be able to be fired in a given line-time, thereby losing one of the major advantages of redundancy.
It would be desirable to provide a method of printing whereby fluctuations in a peak power requirement are minimized. It would be further desirable to provide a method of printing whereby the average power requirement of the printhead is substantially equal to the peak power requirement at any given time during printing. It would be further desirable to provide a method of printing, whereby, in addition minimizing fluctuating peak power requirements, the visual effects of dead or malfunctioning nozzles are reduced. It would be further desirable to provide a method of printing, whereby, in addition to minimizing fluctuating peak power requirements, the visual effects of misdirected ink droplets is reduced.